59
61
# where the basis and destination are unchanged.
61
63
# FIXME: Sometimes we will be given a parents list for a revision
62
# that includes some redundant parents (i.e. already a parent of
63
# something in the list.) We should eliminate them. This can
64
# that includes some redundant parents (i.e. already a parent of
65
# something in the list.) We should eliminate them. This can
64
66
# be done fairly efficiently because the sequence numbers constrain
65
67
# the possible relationships.
67
# FIXME: the conflict markers should be *7* characters
70
from cStringIO import StringIO
73
from bzrlib.lazy_import import lazy_import
74
lazy_import(globals(), """
75
from bzrlib import tsort
81
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
82
RevisionAlreadyPresent,
84
UnavailableRepresentation,
86
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
87
import bzrlib.patiencediff
88
from bzrlib.revision import NULL_REVISION
89
from bzrlib.symbol_versioning import *
71
from difflib import SequenceMatcher
90
73
from bzrlib.trace import mutter
91
from bzrlib.versionedfile import (
98
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
101
class WeaveContentFactory(ContentFactory):
102
"""Content factory for streaming from weaves.
104
:seealso ContentFactory:
107
def __init__(self, version, weave):
108
"""Create a WeaveContentFactory for version from weave."""
109
ContentFactory.__init__(self)
110
self.sha1 = weave.get_sha1s([version])[version]
111
self.key = (version,)
112
parents = weave.get_parent_map([version])[version]
113
self.parents = tuple((parent,) for parent in parents)
114
self.storage_kind = 'fulltext'
117
def get_bytes_as(self, storage_kind):
118
if storage_kind == 'fulltext':
119
return self._weave.get_text(self.key[-1])
120
elif storage_kind == 'chunked':
121
return self._weave.get_lines(self.key[-1])
123
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
126
class Weave(VersionedFile):
74
from bzrlib.errors import WeaveError, WeaveFormatError, WeaveParentMismatch, \
75
WeaveRevisionNotPresent, WeaveRevisionAlreadyPresent
76
from bzrlib.tsort import topo_sort
127
80
"""weave - versioned text file storage.
129
82
A Weave manages versions of line-based text files, keeping track
130
83
of the originating version for each line.
270
196
return self._parents == other._parents \
271
197
and self._weave == other._weave \
272
and self._sha1s == other._sha1s
198
and self._sha1s == other._sha1s
274
201
def __ne__(self, other):
275
202
return not self.__eq__(other)
277
def _idx_to_name(self, version):
278
return self._names[version]
280
def _lookup(self, name):
204
def __contains__(self, name):
205
return self._name_map.has_key(name)
207
def maybe_lookup(self, name_or_index):
208
"""Convert possible symbolic name to index, or pass through indexes."""
209
if isinstance(name_or_index, (int, long)):
212
return self.lookup(name_or_index)
215
def lookup(self, name):
281
216
"""Convert symbolic version name to index."""
282
if not self._allow_reserved:
283
self.check_not_reserved_id(name)
285
218
return self._name_map[name]
287
raise RevisionNotPresent(name, self._weave_name)
220
raise WeaveRevisionNotPresent(name, self)
290
"""See VersionedFile.versions."""
291
223
return self._names[:]
293
def has_version(self, version_id):
294
"""See VersionedFile.has_version."""
295
return (version_id in self._name_map)
297
__contains__ = has_version
299
def get_record_stream(self, versions, ordering, include_delta_closure):
300
"""Get a stream of records for versions.
302
:param versions: The versions to include. Each version is a tuple
304
:param ordering: Either 'unordered' or 'topological'. A topologically
305
sorted stream has compression parents strictly before their
307
:param include_delta_closure: If True then the closure across any
308
compression parents will be included (in the opaque data).
309
:return: An iterator of ContentFactory objects, each of which is only
310
valid until the iterator is advanced.
312
versions = [version[-1] for version in versions]
313
if ordering == 'topological':
314
parents = self.get_parent_map(versions)
315
new_versions = tsort.topo_sort(parents)
316
new_versions.extend(set(versions).difference(set(parents)))
317
versions = new_versions
318
elif ordering == 'groupcompress':
319
parents = self.get_parent_map(versions)
320
new_versions = sort_groupcompress(parents)
321
new_versions.extend(set(versions).difference(set(parents)))
322
versions = new_versions
323
for version in versions:
325
yield WeaveContentFactory(version, self)
327
yield AbsentContentFactory((version,))
329
def get_parent_map(self, version_ids):
330
"""See VersionedFile.get_parent_map."""
332
for version_id in version_ids:
333
if version_id == NULL_REVISION:
338
map(self._idx_to_name,
339
self._parents[self._lookup(version_id)]))
340
except RevisionNotPresent:
342
result[version_id] = parents
345
def get_parents_with_ghosts(self, version_id):
346
raise NotImplementedError(self.get_parents_with_ghosts)
348
def insert_record_stream(self, stream):
349
"""Insert a record stream into this versioned file.
351
:param stream: A stream of records to insert.
353
:seealso VersionedFile.get_record_stream:
356
for record in stream:
357
# Raise an error when a record is missing.
358
if record.storage_kind == 'absent':
359
raise RevisionNotPresent([record.key[0]], self)
360
# adapt to non-tuple interface
361
parents = [parent[0] for parent in record.parents]
362
if (record.storage_kind == 'fulltext'
363
or record.storage_kind == 'chunked'):
364
self.add_lines(record.key[0], parents,
365
osutils.chunks_to_lines(record.get_bytes_as('chunked')))
367
adapter_key = record.storage_kind, 'fulltext'
369
adapter = adapters[adapter_key]
371
adapter_factory = adapter_registry.get(adapter_key)
372
adapter = adapter_factory(self)
373
adapters[adapter_key] = adapter
374
lines = split_lines(adapter.get_bytes(record))
376
self.add_lines(record.key[0], parents, lines)
377
except RevisionAlreadyPresent:
225
def iter_names(self):
226
"""Yield a list of all names in this weave."""
227
return iter(self._names)
229
def idx_to_name(self, version):
230
return self._names[version]
380
232
def _check_repeated_add(self, name, parents, text, sha1):
381
233
"""Check that a duplicated add is OK.
383
235
If it is, return the (old) index; otherwise raise an exception.
385
idx = self._lookup(name)
237
idx = self.lookup(name)
386
238
if sorted(self._parents[idx]) != sorted(parents) \
387
239
or sha1 != self._sha1s[idx]:
388
raise RevisionAlreadyPresent(name, self._weave_name)
240
raise WeaveRevisionAlreadyPresent(name, self)
391
def _add_lines(self, version_id, parents, lines, parent_texts,
392
left_matching_blocks, nostore_sha, random_id, check_content):
393
"""See VersionedFile.add_lines."""
394
idx = self._add(version_id, lines, map(self._lookup, parents),
395
nostore_sha=nostore_sha)
396
return sha_strings(lines), sum(map(len, lines)), idx
398
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
243
def add(self, name, parents, text, sha1=None):
399
244
"""Add a single text on top of the weave.
401
246
Returns the index number of the newly added version.
404
249
Symbolic name for this version.
405
250
(Typically the revision-id of the revision that added it.)
406
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
409
253
List or set of direct parent version numbers.
412
256
Sequence of lines to be added in the new version.
414
:param nostore_sha: See VersionedFile.add_lines.
258
sha -- SHA-1 of the file, if known. This is trusted to be
416
self._check_lines_not_unicode(lines)
417
self._check_lines_are_lines(lines)
419
sha1 = sha_strings(lines)
420
if sha1 == nostore_sha:
421
raise errors.ExistingContent
422
if version_id is None:
423
version_id = "sha1:" + sha1
424
if version_id in self._name_map:
425
return self._check_repeated_add(version_id, parents, lines, sha1)
261
from bzrlib.osutils import sha_strings
263
assert isinstance(name, basestring)
265
sha1 = sha_strings(text)
266
if name in self._name_map:
267
return self._check_repeated_add(name, parents, text, sha1)
269
parents = map(self.maybe_lookup, parents)
427
270
self._check_versions(parents)
428
## self._check_lines(lines)
271
## self._check_lines(text)
429
272
new_version = len(self._parents)
431
275
# if we abort after here the (in-memory) weave will be corrupt because only
432
276
# some fields are updated
433
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
434
# - Robert Collins 20060226
435
277
self._parents.append(parents[:])
436
278
self._sha1s.append(sha1)
437
self._names.append(version_id)
438
self._name_map[version_id] = new_version
279
self._names.append(name)
280
self._name_map[name] = new_version
442
284
# special case; adding with no parents revision; can do
443
285
# this more quickly by just appending unconditionally.
444
286
# even more specially, if we're adding an empty text we
445
287
# need do nothing at all.
447
289
self._weave.append(('{', new_version))
448
self._weave.extend(lines)
290
self._weave.extend(text)
449
291
self._weave.append(('}', None))
450
293
return new_version
452
295
if len(parents) == 1:
554
433
except IndexError:
555
434
raise IndexError("invalid version number %r" % i)
557
def _compatible_parents(self, my_parents, other_parents):
558
"""During join check that other_parents are joinable with my_parents.
560
Joinable is defined as 'is a subset of' - supersets may require
561
regeneration of diffs, but subsets do not.
563
return len(other_parents.difference(my_parents)) == 0
565
def annotate(self, version_id):
566
"""Return a list of (version-id, line) tuples for version_id.
437
def annotate(self, name_or_index):
438
return list(self.annotate_iter(name_or_index))
441
def annotate_iter(self, name_or_index):
442
"""Yield list of (index-id, line) pairs for the specified version.
568
444
The index indicates when the line originated in the weave."""
569
incls = [self._lookup(version_id)]
570
return [(self._idx_to_name(origin), text) for origin, lineno, text in
571
self._extract(incls)]
573
def iter_lines_added_or_present_in_versions(self, version_ids=None,
575
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
576
if version_ids is None:
577
version_ids = self.versions()
578
version_ids = set(version_ids)
579
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
580
if inserted not in version_ids: continue
582
yield line + '\n', inserted
586
def _walk_internal(self, version_ids=None):
587
"""Helper method for weave actions."""
445
incls = [self.maybe_lookup(name_or_index)]
446
for origin, lineno, text in self._extract(incls):
453
(lineno, insert, deletes, text)
454
for each literal line.
592
460
lineno = 0 # line of weave, 0-based
594
462
for l in self._weave:
595
if l.__class__ == tuple:
463
if isinstance(l, tuple):
599
istack.append(self._names[v])
603
dset.add(self._names[v])
605
dset.remove(self._names[v])
607
476
raise WeaveFormatError('unexpected instruction %r' % v)
609
yield lineno, istack[-1], frozenset(dset), l
478
assert isinstance(l, basestring)
480
yield lineno, istack[-1], dset, l
485
def _extract(self, versions):
486
"""Yield annotation of lines in included set.
488
Yields a sequence of tuples (origin, lineno, text), where
489
origin is the origin version, lineno the index in the weave,
490
and text the text of the line.
492
The set typically but not necessarily corresponds to a version.
495
if not isinstance(i, int):
498
included = self.inclusions(versions)
503
lineno = 0 # line of weave, 0-based
509
WFE = WeaveFormatError
511
for l in self._weave:
512
if isinstance(l, tuple):
516
assert v not in istack
530
assert isinstance(l, basestring)
532
isactive = (not dset) and istack and (istack[-1] in included)
534
result.append((istack[-1], lineno, l))
613
537
raise WeaveFormatError("unclosed insertion blocks "
614
538
"at end of weave: %s" % istack)
616
540
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
545
def get_iter(self, name_or_index):
546
"""Yield lines for the specified version."""
547
incls = [self.maybe_lookup(name_or_index)]
548
for origin, lineno, line in self._extract(incls):
552
def get_text(self, name_or_index):
553
return ''.join(self.get_iter(name_or_index))
554
assert isinstance(version, int)
557
def get_lines(self, name_or_index):
558
return list(self.get_iter(name_or_index))
564
def mash_iter(self, included):
565
"""Return composed version of multiple included versions."""
566
included = map(self.maybe_lookup, included)
567
for origin, lineno, text in self._extract(included):
571
def dump(self, to_file):
572
from pprint import pprint
573
print >>to_file, "Weave._weave = ",
574
pprint(self._weave, to_file)
575
print >>to_file, "Weave._parents = ",
576
pprint(self._parents, to_file)
580
def numversions(self):
581
l = len(self._parents)
582
assert l == len(self._sha1s)
587
return self.numversions()
590
def check(self, progress_bar=None):
591
# check no circular inclusions
592
for version in range(self.numversions()):
593
inclusions = list(self._parents[version])
596
if inclusions[-1] >= version:
597
raise WeaveFormatError("invalid included version %d for index %d"
598
% (inclusions[-1], version))
600
# try extracting all versions; this is a bit slow and parallel
601
# extraction could be used
602
nv = self.numversions()
603
for version in range(nv):
605
progress_bar.update('checking text', version, nv)
607
for l in self.get_iter(version):
610
expected = self._sha1s[version]
612
raise WeaveError("mismatched sha1 for version %d; "
613
"got %s, expected %s"
614
% (version, hd, expected))
616
# TODO: check insertions are properly nested, that there are
617
# no lines outside of insertion blocks, that deletions are
618
# properly paired, etc.
622
def merge(self, merge_versions):
623
"""Automerge and mark conflicts between versions.
625
This returns a sequence, each entry describing alternatives
626
for a chunk of the file. Each of the alternatives is given as
629
If there is a chunk of the file where there's no diagreement,
630
only one alternative is given.
632
# approach: find the included versions common to all the
634
raise NotImplementedError()
638
def _delta(self, included, lines):
639
"""Return changes from basis to new revision.
641
The old text for comparison is the union of included revisions.
643
This is used in inserting a new text.
645
Delta is returned as a sequence of
646
(weave1, weave2, newlines).
648
This indicates that weave1:weave2 of the old weave should be
649
replaced by the sequence of lines in newlines. Note that
650
these line numbers are positions in the total weave and don't
651
correspond to the lines in any extracted version, or even the
652
extracted union of included versions.
654
If line1=line2, this is a pure insert; if newlines=[] this is a
655
pure delete. (Similar to difflib.)
657
raise NotImplementedError()
619
660
def plan_merge(self, ver_a, ver_b):
620
661
"""Return pseudo-annotation indicating how the two versions merge.
660
701
# not in either revision
661
702
yield 'irrelevant', line
663
def _extract(self, versions):
664
"""Yield annotation of lines in included set.
666
Yields a sequence of tuples (origin, lineno, text), where
667
origin is the origin version, lineno the index in the weave,
668
and text the text of the line.
670
The set typically but not necessarily corresponds to a version.
673
if not isinstance(i, int):
676
included = self._inclusions(versions)
682
lineno = 0 # line of weave, 0-based
688
WFE = WeaveFormatError
691
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
692
# +285282 0 1676.8040 1676.8040 +<isinstance>
693
# 1.6 seconds in 'isinstance'.
694
# changing the first isinstance:
695
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
696
# +140414 0 762.8050 762.8050 +<isinstance>
697
# note that the inline time actually dropped (less function calls)
698
# and total processing time was halved.
699
# we're still spending ~1/4 of the method in isinstance though.
700
# so lets hard code the acceptable string classes we expect:
701
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
702
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
704
# yay, down to ~1/4 the initial extract time, and our inline time
705
# has shrunk again, with isinstance no longer dominating.
706
# tweaking the stack inclusion test to use a set gives:
707
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
708
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
710
# - a 5% win, or possibly just noise. However with large istacks that
711
# 'in' test could dominate, so I'm leaving this change in place -
712
# when its fast enough to consider profiling big datasets we can review.
717
for l in self._weave:
718
if l.__class__ == tuple:
725
iset.remove(istack.pop())
704
yield 'unchanged', '' # terminator
708
def weave_merge(self, plan):
713
for state, line in plan:
714
if state == 'unchanged' or state == 'killed-both':
715
# resync and flush queued conflicts changes if any
716
if not lines_a and not lines_b:
718
elif ch_a and not ch_b:
720
for l in lines_a: yield l
721
elif ch_b and not ch_a:
722
for l in lines_b: yield l
723
elif lines_a == lines_b:
724
for l in lines_a: yield l
733
raise AssertionError()
727
for l in lines_a: yield l
729
for l in lines_b: yield l
736
if state == 'unchanged':
739
elif state == 'killed-a':
742
elif state == 'killed-b':
745
elif state == 'new-a':
748
elif state == 'new-b':
736
isactive = (not dset) and istack and (istack[-1] in included)
738
result.append((istack[-1], lineno, l))
741
raise WeaveFormatError("unclosed insertion blocks "
742
"at end of weave: %s" % istack)
744
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
748
def _maybe_lookup(self, name_or_index):
749
"""Convert possible symbolic name to index, or pass through indexes.
752
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
757
def join(self, other):
758
"""Integrate versions from other into this weave.
760
The resulting weave contains all the history of both weaves;
761
any version you could retrieve from either self or other can be
762
retrieved from self after this call.
764
It is illegal for the two weaves to contain different values
765
or different parents for any version. See also reweave().
753
if isinstance(name_or_index, (int, long)):
756
return self._lookup(name_or_index)
758
def get_lines(self, version_id):
759
"""See VersionedFile.get_lines()."""
760
int_index = self._maybe_lookup(version_id)
761
result = [line for (origin, lineno, line) in self._extract([int_index])]
762
expected_sha1 = self._sha1s[int_index]
763
measured_sha1 = sha_strings(result)
764
if measured_sha1 != expected_sha1:
765
raise errors.WeaveInvalidChecksum(
766
'file %s, revision %s, expected: %s, measured %s'
767
% (self._weave_name, version_id,
768
expected_sha1, measured_sha1))
771
def get_sha1s(self, version_ids):
772
"""See VersionedFile.get_sha1s()."""
774
for v in version_ids:
775
result[v] = self._sha1s[self._lookup(v)]
778
def num_versions(self):
779
"""How many versions are in this weave?"""
780
l = len(self._parents)
783
__len__ = num_versions
785
def check(self, progress_bar=None):
786
# TODO evaluate performance hit of using string sets in this routine.
787
# TODO: check no circular inclusions
788
# TODO: create a nested progress bar
789
for version in range(self.num_versions()):
790
inclusions = list(self._parents[version])
793
if inclusions[-1] >= version:
794
raise WeaveFormatError("invalid included version %d for index %d"
795
% (inclusions[-1], version))
797
# try extracting all versions; parallel extraction is used
798
nv = self.num_versions()
803
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
804
# The problem is that set membership is much more expensive
805
name = self._idx_to_name(i)
808
new_inc = set([name])
809
for p in self._parents[i]:
810
new_inc.update(inclusions[self._idx_to_name(p)])
812
if set(new_inc) != set(self.get_ancestry(name)):
813
raise AssertionError(
815
% (set(new_inc), set(self.get_ancestry(name))))
816
inclusions[name] = new_inc
818
nlines = len(self._weave)
820
update_text = 'checking weave'
822
short_name = os.path.basename(self._weave_name)
823
update_text = 'checking %s' % (short_name,)
824
update_text = update_text[:25]
826
for lineno, insert, deleteset, line in self._walk_internal():
828
progress_bar.update(update_text, lineno, nlines)
830
for name, name_inclusions in inclusions.items():
831
# The active inclusion must be an ancestor,
832
# and no ancestors must have deleted this line,
833
# because we don't support resurrection.
834
if (insert in name_inclusions) and not (deleteset & name_inclusions):
835
sha1s[name].update(line)
838
version = self._idx_to_name(i)
839
hd = sha1s[version].hexdigest()
840
expected = self._sha1s[i]
842
raise errors.WeaveInvalidChecksum(
843
"mismatched sha1 for version %s: "
844
"got %s, expected %s"
845
% (version, hd, expected))
847
# TODO: check insertions are properly nested, that there are
848
# no lines outside of insertion blocks, that deletions are
849
# properly paired, etc.
767
if other.numversions() == 0:
768
return # nothing to update, easy
769
# two loops so that we do not change ourselves before verifying it
771
# work through in index order to make sure we get all dependencies
772
for other_idx, name in enumerate(other._names):
773
if self._check_version_consistent(other, other_idx, name):
775
for other_idx, name in enumerate(other._names):
776
# TODO: If all the parents of the other version are already
777
# present then we can avoid some work by just taking the delta
778
# and adjusting the offsets.
779
new_parents = self._imported_parents(other, other_idx)
780
lines = other.get_lines(other_idx)
781
sha1 = other._sha1s[other_idx]
782
self.add(name, new_parents, lines, sha1)
851
785
def _imported_parents(self, other, other_idx):
852
786
"""Return list of parents in self corresponding to indexes in other."""
854
788
for parent_idx in other._parents[other_idx]:
855
789
parent_name = other._names[parent_idx]
856
if parent_name not in self._name_map:
790
if parent_name not in self._names:
857
791
# should not be possible
858
raise WeaveError("missing parent {%s} of {%s} in %r"
792
raise WeaveError("missing parent {%s} of {%s} in %r"
859
793
% (parent_name, other._name_map[other_idx], self))
860
794
new_parents.append(self._name_map[parent_name])
861
795
return new_parents
891
def _reweave(self, other, pb, msg):
892
"""Reweave self with other - internal helper for join().
894
:param other: The other weave to merge
895
:param pb: An optional progress bar, indicating how far done we are
896
:param msg: An optional message for the progress
898
new_weave = _reweave(self, other, pb=pb, msg=msg)
899
self._copy_weave_content(new_weave)
901
def _copy_weave_content(self, otherweave):
902
"""adsorb the content from otherweave."""
829
def reweave(self, other):
830
"""Reweave self with other."""
831
new_weave = reweave(self, other)
903
832
for attr in self.__slots__:
904
if attr != '_weave_name':
905
setattr(self, attr, copy(getattr(otherweave, attr)))
908
class WeaveFile(Weave):
909
"""A WeaveFile represents a Weave on disk and writes on change."""
911
WEAVE_SUFFIX = '.weave'
913
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
914
"""Create a WeaveFile.
916
:param create: If not True, only open an existing knit.
918
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
919
allow_reserved=False)
920
self._transport = transport
921
self._filemode = filemode
923
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
924
except errors.NoSuchFile:
930
def _add_lines(self, version_id, parents, lines, parent_texts,
931
left_matching_blocks, nostore_sha, random_id, check_content):
932
"""Add a version and save the weave."""
933
self.check_not_reserved_id(version_id)
934
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
935
parent_texts, left_matching_blocks, nostore_sha, random_id,
940
def copy_to(self, name, transport):
941
"""See VersionedFile.copy_to()."""
942
# as we are all in memory always, just serialise to the new place.
944
write_weave_v5(self, sio)
946
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
949
"""Save the weave."""
950
self._check_write_ok()
952
write_weave_v5(self, sio)
954
bytes = sio.getvalue()
955
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
957
self._transport.put_bytes(path, bytes, self._filemode)
958
except errors.NoSuchFile:
959
self._transport.mkdir(dirname(path))
960
self._transport.put_bytes(path, bytes, self._filemode)
964
"""See VersionedFile.get_suffixes()."""
965
return [WeaveFile.WEAVE_SUFFIX]
967
def insert_record_stream(self, stream):
968
super(WeaveFile, self).insert_record_stream(stream)
972
def _reweave(wa, wb, pb=None, msg=None):
833
setattr(self, attr, getattr(new_weave, attr))
973
837
"""Combine two weaves and return the result.
975
This works even if a revision R has different parents in
839
This works even if a revision R has different parents in
976
840
wa and wb. In the resulting weave all the parents are given.
978
This is done by just building up a new weave, maintaining ordering
842
This is done by just building up a new weave, maintaining ordering
979
843
of the versions in the two inputs. More efficient approaches
980
might be possible but it should only be necessary to do
981
this operation rarely, when a new previously ghost version is
844
might be possible but it should only be necessary to do
845
this operation rarely, when a new previously ghost version is
984
:param pb: An optional progress bar, indicating how far done we are
985
:param msg: An optional message for the progress
989
queue_a = range(wa.num_versions())
990
queue_b = range(wb.num_versions())
850
queue_a = range(wa.numversions())
851
queue_b = range(wb.numversions())
991
852
# first determine combined parents of all versions
992
853
# map from version name -> all parent names
993
854
combined_parents = _reweave_parent_graphs(wa, wb)
994
855
mutter("combined parents: %r", combined_parents)
995
order = tsort.topo_sort(combined_parents.iteritems())
856
order = topo_sort(combined_parents.iteritems())
996
857
mutter("order to reweave: %r", order)
1001
for idx, name in enumerate(order):
1003
pb.update(msg, idx, len(order))
1004
859
if name in wa._name_map:
1005
860
lines = wa.get_lines(name)
1006
861
if name in wb._name_map:
1007
lines_b = wb.get_lines(name)
1008
if lines != lines_b:
1009
mutter('Weaves differ on content. rev_id {%s}', name)
1010
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1012
lines = list(difflib.unified_diff(lines, lines_b,
1013
wa._weave_name, wb._weave_name))
1014
mutter('lines:\n%s', ''.join(lines))
1015
raise errors.WeaveTextDiffers(name, wa, wb)
862
assert lines == wb.get_lines(name)
1017
864
lines = wb.get_lines(name)
1018
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
865
wr.add(name, combined_parents[name], lines)
1022
869
def _reweave_parent_graphs(wa, wb):
1023
870
"""Return combined parent ancestry for two weaves.
1025
872
Returned as a list of (version_name, set(parent_names))"""
1027
874
for weave in [wa, wb]:
1028
875
for idx, name in enumerate(weave._names):
1029
876
p = combined.setdefault(name, set())
1030
p.update(map(weave._idx_to_name, weave._parents[idx]))
877
p.update(map(weave.idx_to_name, weave._parents[idx]))
882
"""Show the weave's table-of-contents"""
883
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
884
for i in (6, 50, 10, 10):
887
for i in range(w.numversions()):
890
parent_str = ' '.join(map(str, w._parents[i]))
891
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
895
def weave_stats(weave_file, pb):
896
from bzrlib.weavefile import read_weave
898
wf = file(weave_file, 'rb')
900
# FIXME: doesn't work on pipes
901
weave_size = wf.tell()
905
for i in range(vers):
906
pb.update('checking sizes', i, vers)
907
for origin, lineno, line in w._extract([i]):
912
print 'versions %9d' % vers
913
print 'weave file %9d bytes' % weave_size
914
print 'total contents %9d bytes' % total
915
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
918
print 'average size %9d bytes' % avg
919
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
923
print """bzr weave tool
925
Experimental tool for weave algorithm.
929
Create an empty weave file
930
weave get WEAVEFILE VERSION
931
Write out specified version.
932
weave check WEAVEFILE
933
Check consistency of all versions.
935
Display table of contents.
936
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
937
Add NEWTEXT, with specified parent versions.
938
weave annotate WEAVEFILE VERSION
939
Display origin of each line.
940
weave mash WEAVEFILE VERSION...
941
Display composite of all selected versions.
942
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
943
Auto-merge two versions and display conflicts.
944
weave diff WEAVEFILE VERSION1 VERSION2
945
Show differences between two versions.
949
% weave init foo.weave
951
% weave add foo.weave ver0 < foo.txt
954
(create updated version)
956
% weave get foo.weave 0 | diff -u - foo.txt
957
% weave add foo.weave ver1 0 < foo.txt
960
% weave get foo.weave 0 > foo.txt (create forked version)
962
% weave add foo.weave ver2 0 < foo.txt
965
% weave merge foo.weave 1 2 > foo.txt (merge them)
966
% vi foo.txt (resolve conflicts)
967
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
979
# in case we're run directly from the subdirectory
980
sys.path.append('..')
982
from bzrlib.weavefile import write_weave, read_weave
983
from bzrlib.progress import ProgressBar
998
return read_weave(file(argv[2], 'rb'))
1004
# at the moment, based on everything in the file
1006
parents = map(int, argv[4:])
1007
lines = sys.stdin.readlines()
1008
ver = w.add(name, parents, lines)
1009
write_weave(w, file(argv[2], 'wb'))
1010
print 'added version %r %d' % (name, ver)
1013
if os.path.exists(fn):
1014
raise IOError("file exists")
1016
write_weave(w, file(fn, 'wb'))
1017
elif cmd == 'get': # get one version
1019
sys.stdout.writelines(w.get_iter(int(argv[3])))
1021
elif cmd == 'mash': # get composite
1023
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
1026
from difflib import unified_diff
1029
v1, v2 = map(int, argv[3:5])
1032
diff_gen = unified_diff(lines1, lines2,
1033
'%s version %d' % (fn, v1),
1034
'%s version %d' % (fn, v2))
1035
sys.stdout.writelines(diff_gen)
1037
elif cmd == 'annotate':
1039
# newline is added to all lines regardless; too hard to get
1040
# reasonable formatting otherwise
1042
for origin, text in w.annotate(int(argv[3])):
1043
text = text.rstrip('\r\n')
1045
print ' | %s' % (text)
1047
print '%5d | %s' % (origin, text)
1053
elif cmd == 'stats':
1054
weave_stats(argv[2], ProgressBar())
1056
elif cmd == 'check':
1061
print '%d versions ok' % w.numversions()
1063
elif cmd == 'inclusions':
1065
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1067
elif cmd == 'parents':
1069
print ' '.join(map(str, w._parents[int(argv[3])]))
1071
elif cmd == 'plan-merge':
1073
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1075
print '%14s | %s' % (state, line),
1077
elif cmd == 'merge':
1079
p = w.plan_merge(int(argv[3]), int(argv[4]))
1080
sys.stdout.writelines(w.weave_merge(p))
1082
elif cmd == 'mash-merge':
1088
v1, v2 = map(int, argv[3:5])
1090
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
1092
base_lines = list(w.mash_iter(basis))
1093
a_lines = list(w.get(v1))
1094
b_lines = list(w.get(v2))
1096
from bzrlib.merge3 import Merge3
1097
m3 = Merge3(base_lines, a_lines, b_lines)
1099
name_a = 'version %d' % v1
1100
name_b = 'version %d' % v2
1101
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
1103
raise ValueError('unknown command %r' % cmd)
1107
def profile_main(argv):
1108
import tempfile, hotshot, hotshot.stats
1110
prof_f = tempfile.NamedTemporaryFile()
1112
prof = hotshot.Profile(prof_f.name)
1114
ret = prof.runcall(main, argv)
1117
stats = hotshot.stats.load(prof_f.name)
1119
stats.sort_stats('cumulative')
1120
## XXX: Might like to write to stderr or the trace file instead but
1121
## print_stats seems hardcoded to stdout
1122
stats.print_stats(20)
1127
if __name__ == '__main__':
1129
if '--profile' in sys.argv:
1131
args.remove('--profile')
1132
sys.exit(profile_main(args))
1134
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
